Apache Polaris and Multi-Engine Iceberg Catalogs

This article discusses Apache Polaris, an open source implementation of an Iceberg REST catalog that solves the problem of catalog lock-in in open lakehouse architectures. It explains why open table formats like Iceberg need open catalogs to achieve full data openness, how Polaris implements REST catalog interoperability for engines like Spark, Flink, Trino, and Dremio, and its role in delegated RBAC across multiple engines. The article also covers Polaris as a multi-cloud metadata plane, key considerations before standardizing, and how it supports an open agentic lakehouse. It provides technical insights into catalog openness as a critical layer for modern data infrastructure.